close

Вход

Забыли?

вход по аккаунту

?

1030386ar

код для вставкиСкачать
Document généré le 25 oct. 2017 17:42
Management international
Management international
ESG Impact on Market Performance of Firms:
International Evidence
Jean-Michel Sahut et Hélène Pasquini-Descomps
Résumé de l'article
Volume 19, numéro 2, Hiver 2015
Cette étude examine dans quelle mesure les informations ESG
(Environnement, Social et Gouvernance d’entreprise),
agrégées sous forme de scores, peuvent influencer les
rendements mensuels des actions sur les marchés en Suisse,
aux États-Unis, et au Royaume-Uni, au cours de la période
2007-2011. Nous constatons que la variation de la note globale
ESG n’est significative qu’au Royaume-Uni. Nous montrons
également que les changements dans les sous-catégories de
notes du GRI (à savoir, la gouvernance, l’économie,
l’environnement, le travail, les droits humains, la société et les
produits) présentent un impact faible mais significatif sur la
performance des actions, mais seulement sur des périodes
restreintes ou pour des secteurs limités, qui varient
contextuellement selon les pays. Enfin, notre régression nonparamétrique met en évidence la non-linéarité probable de la
fonction reliant la performance d’une action à ses
changements de score ESG.
URI : id.erudit.org/iderudit/1030386ar
DOI : 10.7202/1030386ar
Aller au sommaire du numéro
Éditeur(s)
HEC Montréal and Université Paris Dauphine
ISSN 1206-1697 (imprimé)
1918-9222 (numérique)
Découvrir la revue
Citer cet article
Jean-Michel Sahut et Hélène Pasquini-Descomps "ESG Impact
on Market Performance of Firms: International Evidence."
Management international 192 (2015): 40–63. DOI :
10.7202/1030386ar
Tous droits réservés © Management international /
International Management / Gestión Internacional,
2015
Ce document est protégé par la loi sur le droit d'auteur. L'utilisation des services
d'Érudit (y compris la reproduction) est assujettie à sa politique d'utilisation que vous
pouvez consulter en ligne. [https://apropos.erudit.org/fr/usagers/politiquedutilisation/]
Cet article est diffusé et préservé par Érudit.
Érudit est un consortium interuniversitaire sans but lucratif composé de l’Université
de Montréal, l’Université Laval et l’Université du Québec à Montréal. Il a pour
mission la promotion et la valorisation de la recherche. www.erudit.org
ESG Impact on Market Performance of Firms:
International Evidence
L’impact des facteurs ESG sur la performance
en bourse des entreprises : incidences internationales
El impacto de ESG en la rentabilidad de mercado
de las empresas
Jean-Michel SAHUTHélène PASQUINI-DESCOMPS
IPAG Business School, Paris
HEG Haute École de gestion de Genève, Switzerland
Résumé
Cette étude examine dans quelle mesure les
informations ESG (Environnement, Social
et Gouvernance d’entreprise), agrégées
sous forme de scores, peuvent influencer les rendements mensuels des actions
sur les marchés en Suisse, aux ÉtatsUnis, et au Royaume-Uni, au cours de la
période 2007-2011. Nous constatons que
la variation de la note globale ESG n’est
significative qu’au Royaume-Uni. Nous
montrons également que les changements
dans les sous-catégories de notes du GRI (à
savoir, la gouvernance, l’économie, l’environnement, le travail, les droits humains,
la société et les produits) présentent un
impact faible mais significatif sur la performance des actions, mais seulement sur des
périodes restreintes ou pour des secteurs
limités, qui varient contextuellement selon
les pays. Enfin, notre régression non-paramétrique met en évidence la non-linéarité
probable de la fonction reliant la performance d’une action à ses changements de
score ESG.
Abstract
This study investigates how news-based
scores in ESG (Environmental, Social, and
corporate Governance) may have influenced the monthly stocks’ market return
in Switzerland, the US, and the UK during the 2007–2011 period. We find that the
variation of the overall ESG score is only
significant in the UK. We also show that
the changes in sub-category ratings of GRI
(namely, governance, economic, environment, labor, human rights, society, and
products) exhibit a small but significant
impact on the stock’s performance during limited periods or on limited sectors,
which varies among the countries. Finally,
our non-parametric kernel regression highlights that the function linking a stock’s
performance to its ESG-score changes is
probably non-linear.
Keywords: ESG, performance, score,
information, stock, GRI
Mots clés : ESG, performance, score,
information, action, GRI
W
e would like to thank Professor Chris Mallin of
Norwich Business School for her review and constructive comments on an earlier version of this paper.
The question of how and why investors take into account
Corporate Social Responsibility (CSR) activities of firms
when making their investment decision is highly relevant
for research on CSR disclosure and CSR investments, as
well as for firms themselves. This study investigates how
news-based scores in environmental, social, and corporate governance (ESG) may have influenced the monthly
stocks’ market return in Switzerland, the US, and the UK
Resumen
Este estudio investiga cómo​las noticias
basadas en puntuaciones en ESG​(Enviromental, Social & Governance Index)
pueden haber influido​en la​rentabilidad​
mensual de los valores​de las empresas​en
Suiza, los EE.UU. y el Reino Unido durante
el período 2007-2011. ​Los resultados muestran que la variación de la puntuación globa​​
l​en​ESG sólo es significativ​a en el Reino
Unido. ​Además los resultados ​también​
muestran que los cambios​de clasificación
en las subcategorías​(​
es decir, l​​
a gobernanza, ​la ​econ​omía, ​el ​medio ambiente, ​el​
trabajo, ​los ​derechos humanos, ​la ​sociedad
y ​los ​productos) presenta un pequeño pero
significativo impacto en el rendimiento de
las acciones durante períodos limitados
o en sectores limitados,​​​los cuales varían
en los distintos países. Por último, nuestra
regresión ​​no paramétrica​ Kernel​ ​subraya
que la función ​que relaciona el rendimiento
en bolsa con los cambios en las puntuaciones ESG es probablemente no lineal.
Palabras claves: ESG, rendimiento, puntuación, información, acciones, GRI
during the 2007–2011 period. Our model is a multifactor
linear model, consisting of the classic four-factors (FamaFrench’s three factors and momentum), plus a fifth factor,
the ESG score, which represents the potential of the ESG
to explain monthly returns during the observed period. By
linear regression, we find that the variation of the overall
ESG score is not significant in the US and Switzerland for
the observed stocks. In the UK however, the change in the
overall ESG score is a significant and slightly negative
factor of the observed stocks’ monthly performance in the
2007–2010 period. Using the same model, we also study
ESG Impact on Market Performance of Firms: International Evidence
if the changes in sub-categories of ESG ratings (namely,
governance, economic, environment, labor, human rights,
society, and products) could explain the monthly market
return. We find that the changes in sub-category ratings
exhibit a small but significant impact on the stock’s performance during limited periods or on limited sectors, which
varies among the countries. Finally, to explore a possible
non-linear influence of the ESG score over monthly returns,
we use a non-parametric model for Switzerland during the
2007–2011 period. The non-parametric kernel regression
shows that the function linking a stock’s performance to its
ESG-score changes is probably non-linear.
Introduction
Socially responsible investment (SRI) consists of introducing criteria related to sustainability into investment decisions, in contrast to classic investment that focuses solely
on financial criteria. Sustainability criteria are usually organized around three themes: environmental, social/society
and corporate governance (ESG). The first form of SRI,
the exclusion of certain sectors such as weapons, alcohol,
and tobacco for religious or moral purposes, can be traced
back to the 18th century. The exclusion-based strategies
now incorporate exclusions based on recent international
standards and norms and still apply to more than half of
SRI in Europe. In addition, the modern form of SRI uses
various positive screening strategies such as the “best-inclass” approach, which favors companies with better rates,
according to ESG criteria, than other companies in the same
sector (Cf. Appendix A). In addition, active strategies such
as sustainability-themed funds or shareholder rights usage
to direct a corporate strategy are also growing in popularity.
SRI in all its forms has experienced growing popularity
in the last decade1. This interest comes mainly from institutional investors, as public funds undergo further moral
pressure toward sustainability from communities and legislators. The popularity of responsible investment has grown
even more following the 2007 financial crisis that shattered the confidence of investors in financial markets and
traditional investments, while triggering many new policies and rules. SRI proved to be a safer investment during
dropping markets, while rewarding investors with a certain
moral satisfaction, thus emerging as a seductive alternative
investment portfolio approach. It is still unclear, however,
how ESG criteria will reflect into to a firm’s market performance, which is the main question of this study. The question of how and why investors take into account Corporate
Social Responsibility (CSR) activities of firms when making their investment decision is highly relevant for research
1. According to US SIF, assets under SRI strategies went from $2.1
bn in 1999 to $3.7 bn in 2002. EURO SIF claims a 1.7 € bn in 2005,
coming to 11.7 € bn in 2011 which includes norm-based screening since
2009.
41
on CSR disclosure and CSR investments as well as for
firms themselves.
The academic world has been actively studying the field
of modern SRI since the 1990s. This long lasting interest
has been fuelled by the growth in SRI and a lack of a clear
consensus despite numerous studies. Historically, evaluation of SRI studies was hindered by a lack of theory, data,
and methodology (McWilliams and Siegel; 1999, Margolis
et al.; 2007). Recently, ESG-related data have become more
accessible and standardized, and successful methodologies
have been identified. As a result, more and more papers
offering sound theoretical framework as well as strong
associated results are being published, mainly focused on
the American market. However, given large variations in
the empirical results, some authors warn that there is no
conclusive evidence regarding the relationship between
ESG and financial performance of companies (Ioannou and
Serafeim; 2011, Orlitzky; 2013).Therefore, our research
question is how the individual company’s market and financial performance relate to ESG criteria.
The last financial crisis showed the SRI potential to
reduce the risk of an investment through better long-term
management of a company, and this perspective seems
more and more attractive to investors. Our hypothesis is
that companies with high ESG scores have a lower residual
risk and therefore a higher financial performance. We also
believe that publicly available ESG information should
reflect positively in the market price as investors may associate this information with lower residual risk and higher
goodwill.
We therefore propose an original econometric study of
the monthly market performance related to ESG criteria
for major companies in Switzerland; the US, and the UK
between 2007 and 2011.2 Our approach, in order to include
ESG into a company’s market price, is a linear model
using Carhart four-factors plus ESG criteria, as well as a
non-parametric model for kernel regression on the same
variables.
Our results show that the variation of the global ESG
score is a significant and slightly negative factor of a stock’s
monthly performance in the UK, but is not significant in the
US or Switzerland. The changes in sub-categories ratings
(for instance, governance, environment, and labor) exhibit
a small, significant influence over the stock’s performance
only during limited periods or on limited sectors, which
varies among the countries. Moreover, the non-parametric
regression shows that the response of market performance
related to ESG is nonlinear.
Our results provide valuable information for asset managers looking to include ESG criteria into their portfolio
2. The UK period is 2007 to 2010 only, as we did not have the fourfactors for the year 2011 at the time of the study.
42
Management international / International Management / Gestión Internacional, 19 (2)
strategy, and for companies to understand the influence of
ESG news–based ratings on their market price. This study
also contributes to the literature on corporate social responsibility by showing how ESG criteria may link to a firm’s
market performance with a new methodological approach.
The non-parametric response of performance to ESG criteria may open a new way of research to better understand the
complexity of this relationship (Orlitzky; 2013).
CSR and financial performance
Academia seeks actively to demonstrate a connection
between the various ESG criteria and financial performance, and an increasing number of studies have been
devoted to this topic over the past ten years. First we
believe it is important to make a distinction between studies on the overall performance of an SRI portfolio or fund
(Renneboog et al.; 2008, and Galema et al.; 2008) and studies on the financial performance of a single firm or stock
related to its ESG efforts.
The first category, i.e. studies comparing the performance of SRI funds to non-SRI funds, does not take
enough into account the SRI funds’ heterogeneity. Indeed,
the practices of fund management significantly differ in the
world (Sandberg et al., 2009). For instance, almost all SRI
funds in the US use negative screening criteria, which is far
from being the case in Europe. In Europe, the best-in-class
approach- where the leading companies with regard to ESG
criteria from all industries are included in the portfolio- is
the norm, and often considered at the cutting-edge of SRI
(Statman and Glushkov, 2009). A few studies try to overcome this heterogeneity. For instance, Capelle-Blancard
and Monjon (2011) use a different approach by looking into
the determinants of the financial performance among the
SRI funds. They demonstrate that a higher screening intensity reduces the risk-adjusted return, but this result is significant only for sector-specific screening criteria; transversal
screening criteria do not necessarily lead to poor diversification, and so, do not reduce financial performances. For
all these reasons, it is not straightforward to associate the
performance of an SRI fund or constructed portfolio to the
performance of its individual stocks as this would require
additional theories on the construction and management
of the portfolio. As a result, we will focus our following
literature review primarily on studies that help explaining
the link between a single firm’s ESG commitment and its
stock’s performance. To begin with, we will look at studies
that explore why ESG could signal a change in the financial performance of a corporate. For a single company, the
stock’s market performance should later adjust to the corporate’s operational and financial performance, at least in
the semi-strong form of the efficient-market hypothesis.
Linking Social Responsibility and Corporate
Performance
Regarding the definition of a “responsible” company, a
theory often mentioned is the stakeholder theory of R.E.
Freeman (1984). His theory of modern management says
that the managers of a company must take into account all
stakeholders, that is to say, employees, civil society, and
suppliers in their investment decisions and not just shareholders. Although the stakeholder theory has laid a framework in the methods of corporate social responsibility (for
instance ISO 26000 on Global Reporting Initiative uses
methods similar to those suggested by Freeman), it does
not provide information about the relative performance of a
company applying ESG principles in relation to its peers. As
a result, several studies tried to identify and evaluate these
effects and show that CSR activities can create opportunities for firms: to increase image or sales (Albuquerque et al.,
2012); to attract or motivate employees (Balakrishnan et al.
2011); to lower the costs of capital (El Ghoul et al. 2011); to
reduce the “residual risk” (Sharfman and Fernando, 2008);
or to anticipate “best practices” (Eccles et al. 2012).
A prevailing view on the positive impact of ESG activities is to enhance a firm’s image—let us call it the “ESG
advertising” effect. From a marketing perspective, adopting
a policy of sustainability would provide costs and benefits
similar to those of an advertising campaign. Waddock and
Graves (1997) demonstrated a strong relationship between a
company’s reputation (according to the list of most admired
by Fortune magazine) and its ratings in social responsibility. The impact of ESG advertising seems bigger for firms
whose clients are individuals rather than other firms. A
survey for Switzerland from Birth et al. (2008) surveyed
the 300 largest Swiss companies on their CSR communication; 81% of respondents claimed to direct their communication toward customers and 62% point out that their
primary objective is customer loyalty. In addition, a recent
work (Albuquerque et al., 2012) demonstrates that ESG is
a strategic product sold to clients by a company and that
this product is bringing more positive revenues the sooner
it is created, with late followers receiving less value from it.
In a similar way, Porter and Kramer (2011) showed
that CSR could become part of a company’s competitive
advantage if approached in a strategic way. In particular,
societal concerns can yield productivity benefits to a company; “society benefits because employees and their family become healthier, and the firm minimizes employee’s
absences and loss of productivity”. Moreover, a global
survey of 1,122 corporate executives suggests that CEOs
perceived benefits from CSR because it increases attractiveness for potential and existing employees (Economist,
2008). The research of Battacharya et al. (2008) and
Balakrishnan et al. (2011) tend to confirm those findings.
The latest, using a laboratory experiment, show how corporate giving to charity motivates employees. They highlight a double effect: a strong altruism effect and a signaling
ESG Impact on Market Performance of Firms: International Evidence
effect. First, employees contribute more to employers as the
level of corporate giving increases, even if their contribution solely goes to charity. Second, even when employees
compete with charity to get back part of their contributions,
employees’ contributions will increase as the level of corporate giving increases: a charitable employer may signal
more sharing of benefits for employees.
Among the reasons why ESG should lead to increased
performance for a firm, a widely accepted theory in SRI is
the “cost of capital” reduction. The prevailing opinion is
that the costs incurred by the establishment of a socially
responsible structure in a company are offset by a decrease
in its cost of capital. In view of this, Mackey et al. (2007)
postulates that responsible behavior is a “product” sold
by companies to socially responsible investors; but is this
product a profitable one for a company? Previous studies
tend to believe that the impact of investors’ opinion on the
cost of capital is not a significant one. Angel and Rivoli
(1997) demonstrated through an analysis based on the
CAPM that the impact of a boycott of shareholders on the
cost of capital of a company would probably be small if
less than 65% of the shareholders were boycotting the firm.
Similarly, Teoh, Welch, and Wazzan’s (1999) study on the
largest shareholder boycott in South Africa shows minimal
impact on securities. With SRI investments reaching about
12% of all institutional investment in the US as of 2010,
this could be a bone of contention. However, a recent analysis from El Ghoul et al. (2011), using accounting models on
American firms, reveals a constantly lower cost of capital
for firms with high SRI ratings (KLD rating), bringing a
renewed interest to the cost of capital theory.
Another common theoretical position around ESG
and firms’ performance is the residual risk’s “information effect.” Several authors (Kurtz, 2005; Sharfman and
Fernando, 2008) argue that the ratings of a company on
non-accounting parameters tell us about how the company
controls the risks it faces. Therefore, high ESG ratings
would mean lower residual risk for such companies compared to the market. This paradigm is tightly linked to the
well-known reputational risk. The media in the last ten years
have evolved tremendously and the propagation of news,
both good and bad, is now extremely fast. A reputation
risk issue on ESG criteria could affect the company market price,3 or even destroy a thus-far successful company.4
The risk reduction effect of ESG is not to be neglected, as
reputation risk arises as a major threat for companies today.
43
anticipation theory claims two type benefits: first, sustainable companies should also have a better distribution of costs
in relation to upgrading to future regulations. This could be
measured, for instance, by the stability of cash flows over
time, in contrast to other companies increased spending to
adapt to new regulations in target years. Secondly, companies putting in place regulations before others are the leaders in best practices, they are more advanced and forward
thinking compared to their peers, which should lead to an
increase in its wealth and the wealth of its shareholders.
This is what Garriga and Melé (2004) call the instrumental
theory of corporate social responsibility, further supported
in a recent paper from Eccles et al. (2012), who explains
from a management standpoint how mandatory innovation
in products, processes, and business models in sustainable
firms leads to better performance.
In contrast, let us now review some theories on how
high ESG standards could negatively affect a firm’s performance. One can reply to the stakeholder theory that the primary purpose of a business is solely to increase the wealth
of its shareholders (Friedman, 1962), and any other purpose
diverting the firm from this purpose will make it less effective. Some work such as Mackey et al. (2007) and Graff,
Zivin, and Small (2005) argue that a shareholder expects
from a firm to maximize its wealth without ESG constraints,
and that ESG engagement should be done separately, by for
instance giving to charity. A shareholder investing in a firm
with ESG constraints makes a consumption choice where
the charity portion is going to the firm, hence he expects a
lower cost of capital from the firm. This model should lead
to neutral effect for the performance of firms with high ESG
ratings, but it does not account for the risk reduction effect
of ESG.
Another branch bringing controversy are the recent
studies on “sin stocks.” Hong and Kacperczyk (2006) and
Statman and Glushkov (2008) studied “sin stocks” (tobacco,
weapons, alcohol) and found that they shows superior performance to the same extent as companies highly praised
by socially responsible investors. Consequently, they argue
that contrary to common belief, social responsibility efforts
as such are not reflected in the share price.
One last group of principles concerns what could be
called the “best practices’ anticipation” theory. Porter (1991)
explains, about environmental regulations, that the costs
arising from the implementation of a sustainable structure
are offset in time by improving business productivity. This
To summarize, setting-up an ESG program within a
firm has some costs that the firm expects to be compensated
by an advertising effect, more stable revenues from loyal
clients and motivated employees, and a possibly lower cost
of capital, i.e., lower expected return from investors. In the
process, the company might as well lower its risk and perform better, because considering all of its stakeholders will
bring a broader view of its risks and processes. Our first
hypothesis is therefore:
3. Apple’s Foxconn scandal on labor conditions may have cause share
prices to drop 5% when it was announced, taking all other factors into
account. http://seekingalpha.com/article/926801-did-foxconn-bringdown-apple-stock
4. Following the Jan.2013 horsemeat scandal, the French company Spanghero filed for bankruptcy in April 2013 http://www.
huffingtonpost.fr/2013/04/19/viande-cheval-spanghero-place-liquidation-judiciaire_n_3115675.html
44
Management international / International Management / Gestión Internacional, 19 (2)
We expect a slightly positive relationship between
yearly ESG ratings of a firm and its yearly financial
performance. (H1.a )
This concept of synergies created within a firm by engaging
with stakeholders, whether it is clients, business partners, or
employees, is not quite new. It could be considered as part
of the goodwill priced on top of the book value by investors.
Therefore, when a positive ESG score or news is published,
we should observe higher demand, growth, and higher market prices for the corresponding firm as investors should
recognize this added value and lower residual risks. This
additional value and lower residual risk should be reflected
in a stocks market model as a positive alpha of the stock.
We expect a slightly positive relationship between
monthly ESG ratings of a firm and its monthly riskadjusted market performance. (H1.b)
This is consistent with the findings of Gompers, Ishii,
and Metrick (2003) who found that low-rated companies
in terms of governance had a risk-adjusted performance
below average. A study by Russo and Fouts (1997) also
showed that, after adjusting for the most probable parameters (size, growth, media, finance, and others), companies
with better environmental scores had a better-than-average
performance. More recently, Edmans (2007) also found,
taking into account the parameters of the model of Carhart
four-factors (market risk, size, style, and momentum) that
companies ranked by Fortune among the one hundred mostdesirable employers outperformed the average.
Finally, a few excellent meta-analyses have been performed on SRI studies that summarize the findings in the
domain and provide a good overview of the methods used.
The synthesis work carried out by Orlitzky et al. (2003)
and more recently, Margolis et al. (2007) for instance, concludes that there is, in general, a slightly positive relationship between ESG and financial performance of companies,
although less so over the last decade. However, given large
variations in the empirical results, some authors warn that
there is no conclusive evidence regarding this correlation
and emphasize that explanations for the link are complex
(Ioannou and Serafeim; 2011, Orlitzky; 2013).
Measuring the financial and CSR performances
Indeed, though the link between a firm’s market performance and ESG criteria has been much discussed in
recent literature, the empirical results, however, are often
inconclusive. This lack of consistency in the results may
be explained by the multiplicity of data and methodologies used among studies. Especially, the strength of the
link between financial and CSR performances depends on
the way the two performances are measured and numerous moderating variables (Gramlich and Finster 2013).
With support from the above-mentioned meta-analyses and
additional ones cited below, we review the methods used in
previous studies leading to significant results and summarize our findings below.
There is no doubt that the model used in the studies
to evaluate a firm’s performance plays a central role. We
can distinguish first between studies that assess the market performance (stock market returns) and the accounting
financial performance of a company. In general, accounting
models more often bring significant, positive results than
market models. An example of an accounting model is the
Ohlson (1995) model with ROE, ROA, and Tobin’s q variables. The major problem with accounting models is the
number of samples, as it is limited to yearly or quarterly
observations that may be hard find for long periods (over
ten years). For market models, the simple CAPM model
has been progressively abandoned in the profit of multifactor models such as Fama and French, Fama and MacBeth
and Carhart (1997) models. Regressions on such multifactor models generally lead to significant positive results,
whereas CAPM-based models bring little results.
Logic would suggest that working on the most recent
practicable data with the longest possible observation period
would provide a certain significance during statistics tests;
however, the availability of ESG data might limit the ability of the researchers. Revelli and Viviani’s (2013) recent
meta-analysis shows that an observation period of less than
5 years tends to show negative coefficients, whereas 5 to
10 years of data usually bring the most positive results.
They also record that having an observation panel of more
than 100 samples will greatly increase the significance.
Nonetheless, the most common practical issue causing
discrepancies in results might be the sampling frequency.
Orlitzky et al. (2003) believe it to be the main cause of variance among studies in corporate social responsibility.
It should be emphasized that each of the three categories
of ESG scores, whether it is environment, society, or governance, brings overall positive results regarding accounting
performance. However, if we speak about market or fund
performance, the results vary greatly with the selected category, which could explain why previous findings argue that
stock market rewards are rarely observable at the aggregate
level. Hence, we can expect, if using a market model, that
ratings in different subcategories could bring a neutral, negative, or a positive influence. Therefore, we add the following hypothesis to our study: Environmental, Social/Society
or Governance factors do not affect market performance
in the same proportion (H2)
The most studied ESG category is by far governance,
whose positive effect brings a consensus among studies
(Orlitzky et al., 2003); second is environment, while society
factors are the less studied. Horváthová’s (2010) meta-analysis on ecological studies warns that a simple correlation
coefficient will bring more negative results when linking performance to ecological factors. Therefore it seems
appropriate to rely on advanced econometric methods
instead. She also warns that a positive link is found more
ESG Impact on Market Performance of Firms: International Evidence
frequently in common law countries than in civil law countries, which bring us to our next topic.
Concerning the country of observation, there seems to
be a difference in the results obtained in the US and other
countries. Studies in the US bring positive results more
often, while non-US studies lead to neutral results. An
attempt to justify these discrepancies is the activism of US
pension funds toward sustainability. An interesting study
would be to compare emerging markets, as well as the influence of the legal system toward ESG results across categories as Horváthová (2010) did, but this can be made difficult
as most data providers focus on developed countries.
To summarize our findings, to provide certain significance during statistics tests, a study should make the choice
of an accounting model or a multifactor market model as
a base for their performance model. If a market model is
used, we should break down the ESG observation into subcategories, as the aggregated score would lead to no result.
The observation period should be over 5 years or at least
100 samples. There might be a need to resample the data
according to previous studies if no significant results can
be found. Finally, we should expect less positive results in
non-US studies that in US ones.
Methodology
Models
We propose below an original study of over 200 large US,
UK, and Swiss companies, based on the availability of ESG
scores and Fama-French factors. Our study on the performance of companies will compare their ESG ratings available from Covalence with their market performance adjusted
for various factors during the 2007–2011 period. We measured the change in the market value of a stock using a fivefactor linear market model derived from Carhart’s model
(Carhart, 1997). Carhart’s model explains a stock’s market
performance based on the Fama-French three factors (Fama
and French, 1993), namely the market’s excess return (RMRF), the small firms’ excess return (SMB), and the growth
firms’ excess return (HML). In addition, Carhart’s fourfactors model adds the momentum factor (WML) to model
the market trend anomaly. Our hypothesis to add our fifth
factor, called ESG, is that the ESG score variations could
explain partly the stocks’ performance, as it would represent the overall opinion of investors about a corporate’s
ability to lower its risks and anticipate trends. We expect
a neutral or slightly positive relationship between ESG ratings and adjusted market performance (Hypothesis H1.b).
(Model 1)
StockReturnt= RFt + a1(RM–RF)t + a2SMBt + a3HMLt
+a4WMLt + a5 EGSt + a0 + et
(StockReturn – [(RF)]t = a1(RM–RF)t + a2SMBt +
a3HMLt +a4WMLt + a5 EGSt + a0 + et
45
with
Stock Return = monthly company stock’s performance
RF = monthly risk free rate
(RM-RF) = monthly performance of the Market Index,
minus RF
SMB = difference in performance between small and large
companies (by market capitalization)
HML = difference in performance between growth and
mature companies
WM L= differential performance between companies with a
positive or negative trend over the past month
ESG = monthly change in ESG overall score or sub-score
see details in section 4-DATA
In addition, we want to test if the relation with each
factor is indeed linear. In case of the four-factors, the wide
recognition of those factors might have shaped the response
in a linear way. However, in case of the ESG score, we
believe that the positive variations or negative variations
may not affect the stocks in the same way, and that the magnitude of the change in ESG score might affect the stock’s
performance in a non-linear way. To test the shape of this
response without constraint, we conduct a non-parametric
regression on the five factors of the first model.
(Model 2)
(StockReturn – [(RF)]t = f1(Rm–Rf)t + f2(SMB)t +
f3(HMLt ) +f4(WMLt )+ f5(EGSt )+ a0 + et
where
f1 to f5 are functions that will be identified during the
regression to minimize the error under constraints.
In parametric regression, we must determine the
functions f(x) from the start. In non-parametric regression, no hypothesis is made about the f(x) functions;
instead, it is deduced from the data themselves. The objective of the kernel regression is to find a non-linear relation i.e., f(x) between two random variables, in our case
(StockReturn-RF) and each other variable of the model.
As in ordinary least squares (OLS), a weighted sum of the
(StockReturn-RF) observations is used to obtain the fitted values. An important parameter when fitting the curve
to observation is the bandwidth, which provides smoothing so that only some level variation will affect the fitting,
and “noise” variation, on the contrary, will not affect it. We
estimate the unknown regression function using NadarayaWatson kernel implemented in the R “np” package that uses
automatic (data-driven) bandwidth selection.
3.2 Dependent and Independent Variables
The stock market return (StockReturn) is computed
monthly for each stock based on month-end close prices by
Telekurs. For Switzerland, the risk-free rate (RF) and four
46
Management international / International Management / Gestión Internacional, 19 (2)
factors (RM-RF, SMB, HMW, and WML) are available
until 2011 on the Amman-Steiner website.5 RF is the Swiss
Franc call money rate from Factset and the market return
is a constructed portfolio bringing returns very similar to
the Swiss performance index (SPI). The UK four-factors
are taken from the University of Exeter’s6 website, available until 2010 at the time of our study. RF (risk-free rate)
is the monthly return on three-month UK Treasury bills,
while RM is the total return computed on the FT All-Share
Index. The four factors for the US are available on the Jason
Hu website7 until June 2011 where RF also represents the
yield of three-month US Treasury bills. More details on the
construction of the factors are available on the respective
websites.
Concerning our ESG variable, it corresponds to the
change in the Global EthicalQuote® score (hereafter global
score or rating) between the beginning and the end of the
observation period. It can also correspond to the change
in each of the respective sub-scores of one the following
sub-category (governance, economic, environment, labor,
human rights, society, products), as we will test those variables successively.
individuals are more impacted by ESG activities (Eccles,
2012), so we want to see if their market prices are differently influenced by ESG news. We also segregate banks and
insurance as a special group because of the indirect influence of the assets holdings.
ESG data
Our first study sample consists of 618 monthly observations
of change in ESG ratings, corresponding market parameters
on 11 stocks for Switzerland from 2007 to 2011. Our second study sample consists of 1,335 monthly observations of
change in ESG ratings and corresponding financial parameters on 32 UK firms, with observation range from year 2007
to 2010. Our last study sample consists of 8,039 monthly
observations of change in ESG ratings and corresponding
financial parameters on 189 US firms, with observations
ranging from 2007 to 2011.
To take into account the specificities of the companies, we
considered two control variables commonly used for the
analysis of results within the same market: firm size and
sector. In our sample, however, the 11 firms are among
medium or large within their respective markets. In a study
on common stock returns, Banz (1981) has shown that
smaller firms have higher returns, but this effect is not distinctive between medium and large firms. Since our sample
only consists of medium and large firms, we tend to believe
that the parameter influencing the stock returns will not
play differently relative to the size factor; therefore, we disregard this factor in our market model.
In each case, the ESG variable corresponds to the
change in the ESG ratings. ESG ratings available nowadays can be categorized as compliance-based ratings and
news-based ratings, this study’s ratings following the second category. The compliance-based ratings depend on
the compliance of a firm with respect to some pre-defined
rules; for instance, CO2 emissions, the presence of external
auditors, the disclosure of a code of business conduct and
ethics. They often follow the Global Reporting Initiative
(GRI) directives, which has set a standard set of rules for
firms to comply with. The rating is then computed depending on how the firm is complying with the rules. Such data
are found, for instance, on Thomson Reuters’s ASSET4 or
CSRHub. The news-based scores, on the other hand, are
based on positive and negative news concerning a company
found in newspapers and other media and which contains
keywords in relation to environment, society, and governance; for instance, trials, charities, and NGO activities.
Regardless of the method chosen to create the ratings,
the awarded ESG scores are classified by most providers
according to large categories of ideals, often in the number
of three (ESG) or four (ecological, corporate governance,
community, i.e., contribution to society, and humanitarian,
i.e., non-operating employees). An overall ESG score that
aggregates all categories is usually available.
Concerning the sector variable, we will split our sample
in the US and UK according to their sectors, as presented in
Table 1. As our sample for Switzerland is too small to consider each sector individually, we decided instead to group
the firms into the three themed groups that are detailed
below. The rationale for the first group is that it seems that
those firms that are selling consumer products directly to
The compliance-based and news–based rating systems
each have certain advantages and disadvantages. The first
method seems easier to assess because it is following a
grid of specific criteria, but the exact knowledge of what is
required to comply with a rule gives companies the freedom
to simulate good conduct by, for instance, disclosing a code
of conduct which is in fact not followed internally. Another
5. http://www.ammannsteiner.ch/
8. Covalence SA is a limited company based in Geneva, Switzerland,
founded in 2001. They provide ESG ratings, news and data of the world’s largest companies to investors, as well as reputation research and
benchmarks to corporations. http://www.covalence.ch/
The Global EthicalQuote® score and the score in each
sub-category are monthly news-based ratings provided by
Covalence8 on various ESG thematic. More details about
how Covalence computes those ratings and how they link
to the Global Reporting Initiative (GRI) are available in our
data section.
Control Variables
6. http://business-school.exeter.ac.uk/research/areas/centres/xfi/
research/famafrench/files/
7. http://www.jasonhsu.org/research-data.html
ESG Impact on Market Performance of Firms: International Evidence
47
Table 1
Sectors of the empirical study
In order to study if ESG (governance, environmental, social, and corporate) scores have a specific impact on a particular sector, the
firms in our study were sorted by sectors. UK and US firms where divided using ICB super-sectors. In Switzerland, as the number of
sample was too small, we grouped the ICB super-sectors in three custom groups by type of activity: Consumer facing, Bank and Insurance, Industry and other.
Sectors for US & UK
Number of companies
Sector groups for Switzerland :
UK
US
Automobiles & Parts
1
4
Banks
5
5
GROUP I - Consumer facing
Basic Resources
3
7
Food & Beverages
Chemicals
6
Personal & Household Goods Compagnie Financiere Richemont SA
Construction & Materials
3
Financial Services
12
GROUP II Banks & Insurance
Banks
Nestlé S.A.
Food & Beverages
3
14
Health Care
2
13
Industrial Goods & Services
1
17
Insurance
Swiss Re AG
Insurance
1
7
Financial Services
Julius Bär Gruppe AG
Media
3
11
Oil & Gas
1
14
GROUP III - Industry & Others
Personal & Household Goods
3
12
Health Care
Retail
4
22
Technology
UBS AG
Credit Suisse Group
Novartis AG
Roche Holding AG
19
Industrial Goods & Services
ABB Ltd.
Telecommunication
2
4
Chemicals
Syngenta AG
Travel & Leisure
1
9
Construction & Materials
Holcim Ltd.
Utilities
2
10
Grand Total
32
189
Total
11
problem is that it offers only a qualitative but not a quantitative appreciation, so it may not allow to compare companies
that both comply with the same criterion. Finally, compliance rules rely on a yearly evaluation, which makes it hard
for re-assessment during the year.
News-based scores have the advantage of being reassessed more often, as they are based on news communicated by the media and may therefore come from
several sources external to the company, providing different opinions in an ad-hoc manner. The major drawback is
the media’s over-exposure of big companies and clientfacing businesses relative to others. Large companies will
be drowned in a flood of accusation by some organizations
or conversely, the media will extensively cover their good
deeds, while smaller companies will remain in the shadows
and often without a realistic score. To address this issue,
advanced news-based scores compute the media exposure
and adjust the ratings accordingly.
Here are more details on how the ESG scores from
Covalence are calculated. The score is obtained by comparing the amounts of positive and negative information collected on the Web, i.e., by subtracting daily the negative
information from the positive information. When a majority
of negative information is observed, the score then becomes
a negative number.
S = score = A - B
With
A = positive information (or ethical bids)
B = negative information (or ethical demands)
To overcome the bias due to media exposure and size, a
rate representing the total volume of information affecting
the company score is introduced into the formula.
Media exposure adjustment:
48
Management international / International Management / Gestión Internacional, 19 (2)
V = volume = A + B R = rate = S / V
Final score = S * R
An erosion factor of 2% per month gives less importance to old news as compared to the latest ones. The final
score takes into account results performed by several human
analysts specialized in ESG.
A text encoded in the database must also be attached
to one or two criteria among the fifty “criteria for business
contribution to human development” listed below. Those
criteria follow the dimensions of the GRI’s sustainability
reporting and are distributed among seven dimensions.
This allows Covalence to compute the sub-score for each
dimension, namely: A_Governance, B_Economic, C_
Environment, D_Labor, E_HumanRights, F_Society, G_
Products. Table 2 summarizes the groups and the criteria
belonging to it.
The availability of sub-ratings in each of the seven ESG
dimensions, on top of the global score, will allow us to test
which group may have an influence on the stock’s excess
return (Hypothesis H2).
Table 2
Methodology of Covalence score
GRI (Global Reporting Initiative) is one of the most renowned standards for sustainability reporting. The news-based scores from Covalence are grouped under seven categories, the GRI dimensions. Each dimension covers specific criteria, which correspondence to the GRI
guidelines G3.1 is provided below. The news-based scores are computed for the seven categories, as well as a Global score that aggregate
all seven dimensions.
Covalence EthicalQuote Criteria © Covalence SA 2012
GRI Dimension
Governance,
Commitments,
and Engagement
Economic
Environmental
GRI Aspect
id
Criteria name
References to GRI G3.1
4. Governance, Commitments,
and Engagement
Governance
1
Governance
United Nations Policy
Commitments to External Initiatives
Stakeholder Engagement
2
3
4
United Nations Policy
Commitments to External
Initiatives
Stakeholder Engagement
Part 2.4
Economic Performance
5
Fiscal Contributions
EC1
Economic Performance
6
Social Sponsorship
EC1
Economic Performance
7
Public Funding
EC4
Market Presence
8
Wages
Market Presence
9
Local Sourcing
EC6
Market Presence
10
Local Hiring
EC7
Indirect Economic Impacts
11
Infrastructures
GRI 3.1 EC8
Indirect Economic Impacts
12
Indirect Economic Impacts
EC9
Indirect Economic Impacts
Indirect Economic Impacts
13
14
Pricing / Needs
Intellectual Property Rights
EC9
EC9
Materials
15
Materials
EN1, EN2
Energy
16
Energy
EN3, EN4, EN5, EN6, EN7
Water
17
Water Management
EN8, EN9, EN10
Biodiversity
18
Biodiversity
EN11, EN12
Emissions, Effluents, and Waste
19
Emissions
EN16, EN17, EN18, EN19, EN20
Emissions, Effluents, and Waste
20
Waste Management
EN21, EN22, EN24, EN25
Emissions, Effluents, and Waste
21
EN23
Products and Services
22
Pollution
Environmental Impacts of
Products
Compliance
23
EN28
Transport
24
Compliance
Environmental Impact of
Transport
EN26, EN27
EN29
ESG Impact on Market Performance of Firms: International Evidence
Labor Practices and
Decent Work
Human Rights
Society
Product
Responsibility
49
Employment
25
Employment
LA1, LA2
Employment
26
Employee Benefits
LA3, LA15
Labor/Management Relations
27
Trade Unions
LA5
Occupational Health and Safety
28
Health and Safety
LA6, LA7, LA8, LA9
Training and Education
29
LA10, LA11, LA12
Diversity and Equal Opportunity
Investment and Procurement
Practices
30
Training and Education
Diversity and Equal Opportunity
31
Human Rights Policy
HR1, HR2, HR3, HR10, HR11
Non-discrimination
32
Discrimination
HR4, LA14
Child Labor
33
Child Labor
HR6
Forced and Compulsory Labor
34
Forced Labor
HR7
Security Practices
Indigenous Rights
35
36
Security Practices
Indigenous Rights
HR8
HR9
Local Communities
37
Local Communities
SO1
Local Communities
38
Humanitarian Action
SO1
Corruption
39
Corruption
SO2, SO3, SO4
Public Policy
40
SO5
Public Policy
41
Lobbying Practices
Contributions to Political
Parties
Anti-Competitive Behavior
42
Competition
SO7
Compliance
43
SO8
Awards
44
Social Compliance
Awards, Reports and Comments
Customer Health and Safety
45
Product Safety
PR2
Product and Service Labeling
46
Product Labeling
PR4
Marketing Communications
47
Marketing Communications
PR6, PR7
Customer Privacy
48
Customer Privacy
PR8
Compliance
Social Impacts of Products
49
50
Product Compliance
Social Impacts of Products
PR9
LA13
SO6
The EthicalQuote index aggregates thousands of documents gathered online from various sources and classified according to 50 sustainability criteria inspired by the Global Reporting Initiative’s G3.1 sustainability reporting guidelines, as well as by the experience
accumulated by Covalence since 2001. These criteria cover the economic, social, environmental and governance impacts of companies.
The Global Reporting Initiative (GRI) is a non-profit organization that promotes economic, environmental and social sustainability. GRI
provides all companies and organizations with a comprehensive sustainability reporting framework that is widely used around the world.
Result
Descriptive Statistics
The descriptive statistics of our first sample (market Model
1 and 2) is summarized in the table below for each country.
For Switzerland, we have 618 observations for each variable over the period 2007–2011 and 8,039 for the US on
the same period. We have 1,335 observations in the UK
between 2007 and 2010. The stock excess returns range
between -53% and +49% in Switzerland, -63% and +90%
in the UK, and -78 and +260% in the US. The ESG ratings
experience a higher range of variations (e.g., Switzerland,
between -4,500% and 180%) than the other dependent variables (e.g., Switzerland, min - 15% and max 12%).
We test positively for normality by drawing histograms,
where the high kurtosis can be noted for the ESG scores.
Heteroscedasticity is tested negatively by using a plot of
each of our independent variables against the square of
the residual, showing no pronounced pattern. The multicolinearity between the Carhart four-factors’ and the
ESG scores’ change is low with VIF indices below 2. The
Pearson correlation between the excess stock return and
the variables are shown in Table 2. In the overall sample,
the four-factors are, as expected, highly correlated with the
stock’s excess return. For Switzerland and the UK stocks
excess return is also correlated to the global ESG score,
positively for Switzerland, and negatively for the UK. The
US does not display any significant correlation between the
stock’s excess return and the global score, but a positive
50
Management international / International Management / Gestión Internacional, 19 (2)
Table 3
Descriptive statistics
Performance Measurement statistics for January 2007 to December 2011. Stock’s return is log-return computed from the monthly market
observations of the close price from Telekurs. RF represents the three -month T-bill return for UK and US, and the call money rate for
Switzerland. RM-RF is the excess return on Fama and French’s (1993) market proxy. SMB and HML are Fama and French’s factors for
size and value. MOM is the Carhart’s factor for momentum. Our ESG factors represent the variation in the news-based score on environmental, social/societal and governance criteria. The “ESG Global Chng” represents the variation in the overall ESG score, while the
seven others “<category> Chng” (for instance ESG A_Governance.Chng) represent the changes in a score computed only in one of the
seven GRI dimension measured by Covalence.
CH 2007-2011
Count
Mean
Standard
Deviation
Kurtosis
Skewness
Minimum Maximum
Range
StockReturn-RF
618
-0.010
0.095
4.70
0.28
-0.53
0.49
1.02
RM-RF
618
-0.004
0.044
0.32
-0.23
-0.11
0.12
0.23
SMB
618
0.001
0.028
0.05
0.21
-0.06
0.09
0.15
HML
618
-0.002
0.022
-0.05
-0.13
-0.06
0.05
0.11
MOM
618
0.005
0.033
6.34
-1.68
-0.15
0.08
0.23
ESG Global Chng
618
0.026
0.717
144.34
6.67
-7.38
11.69
19.08
ESG A_Governance Chng
618
0.001
0.443
113.05
7.36
-2.40
6.56
8.96
ESG B_Economic Chng
618
0.008
0.251
103.46
5.39
-2.16
3.77
5.92
ESG C_Environment Chng
618
0.039
0.395
185.32
12.45
-0.88
6.59
7.47
ESG D_Labor Chng
618
-0.098
1.979
457.44
-19.53
-45.56
12.81
58.37
ESG E_Human Rights Chng
618
0.001
1.318
221.14
-8.06
-24.59
12.83
37.42
ESG F_Society Chng
618
-0.032
1.312
170.82
-3.44
-20.74
18.17
38.91
ESG G_Product Chng
618
-0.015
0.421
415.75
-18.35
-9.48
2.05
11.53
Standard
Deviation
Kurtosis
Skewness
UK 2007-2010
Count
Mean
Minimum Maximum
Range
StockReturn-RF
1’335
0.003
0.107
10.25
0.81
-0.63
0.90
1.53
RM-RF
1’335
0.002
0.054
-0.23
-0.45
-0.14
0.10
0.24
SMB
1’335
-0.002
0.045
5.65
1.15
-0.12
0.19
0.30
HML
1’335
-0.004
0.032
4.60
1.46
-0.07
0.11
0.19
MOM
1’335
0.006
0.064
6.79
-1.85
-0.27
0.14
0.41
ESG Global Chng
1’335
-0.049
1.249
689
-24.34
-38.10
4.68
42.78
ESG A_Governance Chng
1’335
0.018
0.873
490
18.84
-8.75
23.23
31.98
ESG B_Economic Chng
1’335
-0.020
2.890
789
-18.84
-90.39
50.66
141.06
ESG C_Environment Chng
1’335
-0.042
1.315
808
-25.43
-42.19
9.74
51.93
ESG D_Labor Chng
1’335
-0.178
5.149
782
-23.30
-163.34
63.64
226.98
ESG E_Human Rights Chng
1’335
-0.104
2.377
817
-26.86
-76.24
9.84
86.08
ESG F_Society Chng
1’335
-0.318
8.603
947
-29.76
-286.18
26.86
313.04
ESG G_Product Chng
1’335
-0.081
2.377
511
-18.39
-61.94
31.44
93.39
US 2007-2011
Count
Mean
Standard
Deviation
Kurtosis
Skewness
Minimum Maximum
Range
StockReturn-RF
8’039
0.005
0.112
65.74
3.00
-0.78
2.60
3.38
RM-RF
8’039
0.004
0.055
0.48
-0.66
-0.17
0.10
0.27
SMB
8’039
0.005
0.024
-0.36
0.51
-0.03
0.07
0.10
HML
8’039
-0.005
0.039
1.17
0.15
-0.12
0.11
0.22
ESG Impact on Market Performance of Firms: International Evidence
51
MOM
8’039
0.000
0.071
9.22
-2.34
-0.35
0.13
0.48
ESG Global Chng
8’039
0.146
6.615
7’408
84.58
-32.08
581.12
613.20
ESG A_Governance Chng
8’039
0.096
3.986
5’027
65.81
-37.80
316.28
354.08
ESG B_Economic Chng
8’039
0.063
2.376
2’670
41.15
-65.82
157.57
223.38
ESG C_Environment Chng
8’039
0.060
1.860
2’334
43.82
-20.84
103.29
124.13
ESG D_Labor Chng
8’039
0.023
3.473
2’447
38.97
-58.96
226.74
285.70
ESG E_Human Rights Chng
8’039
-0.137
10.824
6’227
-75.32
-909.12
86.73
995.85
ESG F_Society Chng
8’039
-0.012
4.375
5’567
-69.22
-357.37
30.56
387.93
ESG G_Product Chng
8’039
0.097
5.580
7’299
83.54
-41.23
488.38
529.61
one with the labor sub-score changes. Despite the high correlation between the four-factors for all countries and ESG
scores for Switzerland, the VIF indices are low and below
3 for all coefficients.
Model 1 Analysis
We run our regression toward Model 1 in R, with results
presented below. As expected, the market premium RM-RF
shows the highest positive significance toward the stock’s
performance. The other classic factors also display a various degree of significance with an expected negative coefficient for SMB since all of our firms are large-cap and an
expected positive coefficient for our firm since our stocks
are value stocks, confirming global findings on FamaFrench models. The momentum factor seems slightly negative for Switzerland. Our first model linear regression shows
a slightly positive relationship between the EthicalQuote
global score and the market performance; however, it is not
significant. The coefficient factor for the ESG Global score
change over stock’s market performance is 0.004, which is
very small. A bigger sample might be required to confirm
such a small effect in a significant manner.
To explore the influence of each ESG subcategory individually, we then regress for a linear model consisting of four
factors, and the score changes in each of the seven subcategories. The figures are presented below. For Switzerland,
economic news expectedly demonstrates a positive relation
to stock market performance. The overall sample exhibits
a significant negative relation between labor score changes
and the stock’s excess return changes. This small negative
impact of labor ratings over the whole period, which might
confirm Friedman’s (1970) concern that business should
focus on profit only, but this effect tends to disappear in
recent years as we later explore regression by year. Labor
rating results from positive and negative news concerning
the labor practices and decent work, such as employment
and employee benefits, trade unions, health and safety at
work, training and education, and diversity (see Table 2
for equivalent GRI criteria). A bivariate Granger causality
test with a one-period shift shows a highly significant probability that it is the labor’s score change that is causing the
changes in market value. We also consistently measure the
impact of ESG news-based ratings to be smaller in comparison with market premium and smaller than SMB, HML,
and MOM factors.
For the US and UK, only the market premium and
momentum factors show a high degree of significance.
Society news demonstrates a statistically significant relation in the UK over the whole period, but the factor’s coefficient seems too small to be meaningful.
Insert Table 6
In order to further explore the relationship with each
ESG subcategory, we observe each category’s score per
year. For Switzerland, over year 2011, the environment
score exhibits a positive and significant (P < (t) 0.05) influence over the stock market’s performance, while the labor
score’s significant negative coefficient only seems to apply
to the year 2008. Those results suggest that some factors
may be more influential during some periods or context,
as, for instance, 2008’s sensitivity to labor when the financial crisis began. For labor, this could mean that positive
news concerning the employee benefits of employment
are perceived negatively in the markets during a crisis or
more probably that negative news, such as lay-offs, are still
perceived as a positive sign that the business is restructuring, which might be challenged. 2011’s sensitivity to
environmental questions might have been triggered by the
Fukushima Daiichi nuclear disaster or by the 2011 proposal
for a new regulation from the Swiss federal office to cut
CO2 emission, which was finally rejected. The environment
category in our news-based score contains news related to
materials, energy, water management, biodiversity, emission and waste, pollution, ecological impact of products
and transports.
Changes in the society score also show a significant
positive coefficient for year 2008. A bivariate Granger
causality test for each variable with respect to the stock’s
excess return does not enable us to conclude on the direction of causality.
The UK sample demonstrates a negative, but significant, coefficient over the year 2009 for the society score
(local communities, humanitarian action, corruption and
52
Table 4
Correlation between variables
Performance Measurement Model for January 2007 to December 2011. Stock’s return is log-return computed from the monthly market observations of the close price from Telekurs. RF is the
three -month T-bill return for UK and US, and call money rate for Switzerland. RM-RF is the excess return on Fama and French’s (1993) market proxy. SMB and HML are Fama and French’s
factors for size and value. MOM is the Carhart’s factor for momentum. Our ESG factors represent the variation in the news-based score on environmental, social/societal and governance
criteria. The “ESG Global Chng” represents the variation in the overall ESG score, while the seven others “<category> Chng” (for instance ESG A_Governance.Chng) represent the changes
in a score computed only in one of the seven GRI dimension measured by Covalence.
CH
Stock Return-RF
StockReturnRF
100%
RM - RF
HML
SMB
MOM
60%(***) 30%(***) -33%(***) -36%(***)
ESG Global ESG A_GoChng
vernance
Chng
ESG B_
Economic
Chng
ESG C_En- ESG D_La- ESG E_Hu- ESG F_Sovironment
bor Chng man Rights ciety Chng
Chng
Chng
ESG
G_Product
Chng
0%
4%
-3%
-6%
2%
5%
3%
100%
32%
-42%
-50%
7%(.)
2%
0%
-7%
0%
-1%
2%
1%
HML
100%
-33%
-29%
2%
5%
1%
2%
5%
-7%(.)
5%
6%
SMB
100%
19%
5%
-3%
7%(.)
8%(*)
0%
-3%
1%
-1%
MOM
100%
-6%
-3%
3%
1%
3%
4%
-5%
-3%
ESG Global Chng
100%
27%(***)
16%(***)
4%
9%(*)
3%
10%(*)
9%(*)
ESG A_Governance Chng
100%
15%(**)
3%
8%(*)
2%
6%
14%(**)
ESG B_Economic Chng
100%
2%
16%(***)
4%
13%(**)
3%
ESG C_Environment Chng
100%
1%
2%
-16%(***)
2%
ESG D_Labor Chng
ESG E_Human Rights Chng
100%
1%
1%
0%
100%
9%(*)
-1%
ESG F_Society Chng
100%
5%
ESG G_Product Chng
100%
n=618
UK
StockReturn.
RF
RM.RF
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1
Stock Return-RF
RM - RF
HML
SMB
MOM
ESG Global ESG A_GoChng
vernance
Chng
ESG B_
Economic
Chng
ESG C_En- ESG D_La- ESG E_Hu- ESG F_Sovironment
bor Chng man Rights ciety Chng
Chng
Chng
ESG
G_Product
Chng
100%
51%
(***)
39% (***) 23% (***) -30% (***) -15% (***)
-2%
-1%
-2%
3%
-1%
1%
-1%
100%
69% (***) 28% (***) -35% (***)
-2%
-1%
-2%
3%
-3%
0%
0%
51% (***) -57% (***)
HML
100%
SMB
100%
MOM
-4%
-9% (**)
0%
0%
0%
-1%
-1%
-2%
-1%
-68% (***) -10% (**)
2%
4%
3%
-1%
1%
-4%
2%
2%
-2%
-2%
-1%
1%
0%
-1%
100%
12% (***)
Management international / International Management / Gestión Internacional, 19 (2)
7%(.)
RM - RF
100%
2%
2%
6% (*)
1%
1%
11% (***)
1%
ESG A_Governance Chng
100%
1%
1%
-1%
1%
0%
-2%
ESG B_Economic Chng
100%
3%
1%
0%
2%
1%
ESG C_Environment Chng
100%
1%
0%
0%
1%
ESG D_Labor Chng
100%
3%
0%
0%
100%
-1%
0%
ESG E_Human Rights Chng
ESG F_Society Chng
100%
0%
ESG G_Product Chng
100%
MOM
ESG Global
Chng
ESG A_Governance
Chng
ESG E_Human Rights
Chng
ESG F_Society Chng
ESG G_Product Chng
n=1335
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1
US
Stock ReturnRF
RM - RF
HML
SMB
ESG B_Eco- ESG C_Envi- ESG D_Lanomic Chng ronment Chng bor Chng
100%
55% (***)
28% (***)
22% (***)
-33% (***)
-1%
1%
1%
-1%
2% (.)
-1%
-2%
-1%
RM.RF
100%
43% (***)
39% (***)
-48% (***)
-2% (*)
1%
1%
-1%
2% (.)
-1%
0%
-1%
HML
100%
18% (***)
-47% (***)
3% (*)
0%
0%
0%
1%
0%
1%
0%
StockReturn.RF
SMB
100%
-14% (***)
-1%
-2%
-2%
-1%
2% (*)
1%
2%
-2%
MOM
100%
1%
-2%
-1%
0%
1%
2% (.)
-1%
0%
ESG Global Chng
100%
1%
4% (**)
1%
0%
0%
1%
0%
ESG A_Governance Chng
100%
1%
0%
0%
3% (*)
1%
0%
ESG B_Economic Chng
100%
0%
0%
0%
1%
0%
ESG C_Environment Chng
100%
-2% (.)
0%
0%
0%
ESG D_Labor Chng
100%
0%
0%
0%
ESG E_Human Rights Chng
100%
0%
0%
ESG F_Society Chng
100%
0%
ESG G_Product Chng
100%
n=8039
ESG Impact on Market Performance of Firms: International Evidence
ESG Global Chng
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1
53
54
Management international / International Management / Gestión Internacional, 19 (2)
Table 5
Regression results for Model 1
Performance Measurement Model for January 2007 to December 2011. Stock’s return is log-return computed from the monthly market
observations of the close price from Telekurs. RF is the three -month T-bill return or UK and US, and the call money rate for Switzerland.
RM-RF is the excess return on Fama and French’s (1993) market proxy. SMB and HML are Fama and French’s factors for size and value.
MOM is the Carhart’s factor for momentum. The “ESG Global Chng” factor represents the variation in the overall news-based score on
environmental, social/societal and governance criteria
StockReturn-RF ~ RM-RF + SMB + HML + MOM + ESG Global Chng
CH
RM-RF
SMB
HML
MOM
ESG Global Chng
(Intercept)
Estimate
Pr(>|t|)
VIF
0.0881
12.3130
<2e-16
1.6037
0.1215
-2.0100
0.0448
1.2994
-0.244 *
0.411 **
-0.196 .
0.004
-0.004
-0.35965
UK
t-value
1.084 ***
Signif. codes: Pr(>|t|) 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1
Residual standard error: 0.07548 on 612 degrees of freedom
Multiple R-squared: 0.375, Adjusted R-squared: 0.3699
F-statistic: 73.44 on 5 and 612 DF, p-value: <
Residuals:
Min
1Q
-0.04201
Estimate
Std. Error
0.1541
2.6650
0.0079
1.2098
0.1079
-1.8160
0.0699
1.3805
0.0043
0.8310
0.4062
1.0138
0.0031
-1.2970
0.1951
Median
-0.0018
0.0000
3Q
0.0398
Max
0.3060
Std. Error
t value
Pr(>|t|)
VIF
RM-RF
0.969 ***
0.0640
15.1440
< 2e-16
1.9392
SMB
0.024
0.0765
0.3130
0.7543
1.9393
HML
-0.115
0.1239
-0.9300
0.3523
2.6133
MOM
-0.211 ***
0.0564
-3.7330
0.0002
2.1385
ESG Global Chng
-0.010 ***
0.0020
-4.8290
0.0000
1.0160
0.0025
0.2600
0.7949
(Intercept)
0.001
Signif. codes: Pr(>|t|) 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1
Residual standard error: 0.09026 on 1329 degrees of freedom
Multiple R-squared: 0.293, Adjusted R-squared: 0.2904
F-statistic: 110.2 on 5 and 1329 DF, p-value: < 2.2e-16
Residuals:
Min
-0.4927
US
Median
3Q
Max
-0.0438
-0.0005
0.0408
0.7558
Std. Error
t value
Pr(>|t|)
VIF
RM-RF
1.005 ***
0.0241
41.7790
< 2e-16
1.6075
SMB
0.070
0.0471
1.4930
0.1350
1.1836
HML
0.072 *
0.0318
2.2740
0.0230
1.3845
MOM
Estimate
1Q
-0.126 ***
ESG Global Chng
0.000
(Intercept)
0.001
0.0178
-7.0810
0.0000
1.4666
0.0002
0.2830
0.7770
1.0024
0.0011
1.0900
0.2760
Signif. codes: Pr(>|t|) 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1
Residual standard error: 0.09356 on 8033 degrees of freedom
Multiple R-squared: 0.3061, Adjusted R-squared: 0.3057
F-statistic: 708.7 on 5 and 8033 DF, p-value: < 2.2e-16
Residuals:
Min
-0.6242
1Q
Median
3Q
Max
-0.04328
-0.00256
0.0403
2.44116
ESG Impact on Market Performance of Firms: International Evidence
55
Table 6
Regression results for Model 1 - Sub-scores
Performance Measurement Model for January 2007 to December 2011. Stock’s return is log-return computed from the monthly market
observations of the close price from Telekurs. RF is the three -month T-bill return for UK and US, and call money rate for Switzerland.
RM-RF is the excess return on Fama and French’s (1993) market proxy. SMB and HML are Fama and French’s factors for size and value.
MOM is the Carhart’s factor for momentum. Our ESG factors represent the variation in the news-based score on environmental, social/
societal and governance criteria. The seven “<category> Chng” (for instance ESG A_Governance.Chng) factors represent the changes
in a score computed only in one of the seven GRI dimension measured by Covalence.
StockReturn-RF ~ RM-RF + SMB + HML + MOM + ESG A_Governance Chng + ESG B_Economic.Chng + ESG C_Environment.
Chng + ESG D_Labor.Chng + ESG E_Human.Rights.Chng + ESG F_Society.Chng + ESG G_Product.Chng
CH
Estimate
Std.Error
tvalue
Pr(>|t|)
VIF
RM-RF
1.087 ***
0.087765
12.38
<2e-16
1.600978
SMB
-0.255 *
0.121912
-2.089
0.0371
1.313569
HML
0.423 **
0.155375
2.721
0.0067
1.235752
MOM
-0.195 .
0.107964
-1.809
0.0709
1.390157
ESG A_Governance Chng
-0.006
0.007012
-0.823
0.4107
1.051974
ESG B_Economic Chng
0.023 .
0.01251
1.873
0.0615
1.075129
ESG C_Environment Chng
0.003
0.007842
0.323
0.7469
1.046932
ESG D_Labor Chng
-0.003 *
0.001558
-2.112
0.0351
1.034534
ESG E_Human.Rights Chng
0.002
0.002324
0.937
0.3491
1.021631
ESG F_Society Chng
0.002
0.00239
0.661
0.5086
1.06963
ESG G_Product Chng
0.004
0.007288
0.546
0.585
1.02624
(Intercept)
-0.004 0.003087
-1.409
0.1594
Signif. codes: Pr(>|t|) 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1
Residual standard error: 0.07529 on 606 degrees of freedom
Multiple R-squared: 0.3841,
Adjusted R-squared: 0.3729
F-statistic: 34.36 on 11 and 606 DF, p-value: < 2.2e-16
Residuals:
Min
1Q
Median
3Q
Max
-0.35766
-0.04
-0.00118
0.03955
0.30425
UK
Estimate
Std. Error
t value
Pr(>|t|)
VIF
RM-RF
0.959 ***
0.06482
14.801
< 2e-16
1.947188
SMB
0.036
0.07756
0.47
0.639
1.951354
HML
-0.093
0.1254
-0.743
0.458
2.620275
MOM
-0.224 ***
0.05701
-3.922
0.0000925
2.139914
ESG A_Governance Chng
-0.001
0.002865
-0.382
0.703
1.003457
ESG B_Economic Chng
0.000
0.0008656
-0.532
0.595
1.003316
ESG C_Environment Chng
-0.001
0.001902
-0.412
0.68
1.002886
ESG D_Labor Chng
0.000
0.0004861
0.471
0.638
1.004173
ESG E_Human.Rights Chng
0.000
0.001052
-0.089
0.929
1.00196
ESG F_Society Chng
0.000
0.0002909
0.189
0.85
1.003889
ESG G_Product Chng
0.000
0.001051
-0.253
0.8
1.001766
(Intercept)
0.001 0.002561
0.523
0.601
Signif. codes: Pr(>|t|) 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1
Residual standard error: 0.09122 on 1323 degrees of freedom
Multiple R-squared: 0.2811, Adjusted R-squared: 0.2752
F-statistic: 47.04 on 11 and 1323 DF, p-value: < 2.2e-16
Residuals:
Min
1Q
Median
3Q
Max
-0.49155 -0.0438
-0.00108
0.04055
0.74787
56
Management international / International Management / Gestión Internacional, 19 (2)
StockReturn-RF ~ RM-RF + SMB + HML + MOM + ESG A_Governance Chng + ESG B_Economic.Chng + ESG C_Environment.
Chng + ESG D_Labor.Chng + ESG E_Human.Rights.Chng + ESG F_Society.Chng + ESG G_Product.Chng
US
RM-RF
SMB
HML
MOM
ESG A_Governance Chng
ESG B_Economic Chng
ESG C_Environment Chng
ESG D_Labor Chng
ESG E_Human.Rights Chng
ESG F_Society Chng
Estimate
1.004
0.071
0.073
-0.126
0.000
0.000
0.000
0.000
0.000
0.000
***
*
***
.
Std. Error
0.02406
0.0471
0.03182
0.01776
0.000262
0.0004394
0.0005613
0.0003007
0.00009648
0.0002386
t value
41.713
1.516
2.297
-7.108
0.348
0.862
-0.886
0.983
0.272
-1.776
Pr(>|t|)
< 2e-16
0.1295
0.0216
1.28E-12
0.7279
0.3887
0.3759
0.3256
0.7858
0.0758
VIF
1.60753
1.185605
1.382455
1.468276
1.001718
1.000742
1.00075
1.001574
1.001517
1.000549
ESG G_Product Chng
0.000
0.000187
-0.244
0.807
1.000373
(Intercept)
0.001 0.001084
1.091
0.2753
1Q
Median
3Q
Max
-0.0433
-0.0026
0.04035
2.44102
Signif. codes: Pr(>|t|) 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1
Residual sstandard error: 0.09356 on 8027 degrees of freedom
Multiple R-squared: 0.3066, Adjusted R-squared: 0.3057
F-statistic: 322.7 on 11 and 8027 DF, p-value: < 2.2e-16
Residuals:
Min
-0.62426
lobbying, etc.) has a negative significant relation with market performance, which might be a reaction to the lingering recovery and the MP expenses scandal causing defiance
toward anything but economic value. The economic score,
which gathers new related to economic performance and
social factors, such as wages, local sourcing and hiring, and
property rights has a positive significant relation with market performance for 2010, but the causality is not confirmed
by the Granger test. Therefore, it is unknown if the firms
improved their socio-economics because of better performances that usual, or the firms with a higher socio-economic score were having better market performances. The
US sample shows a positive significant coefficient toward
society score changes in 2007 and a slightly negative one
for the year 2009 regarding product changes (product safety
and labeling, product social impact, consumer privacy, etc.),
but both impacts are very small. As we will see later with
our split by sector, products score shows a significant positive relation to the technology sector in the whole period.
As described in our methodology section, we then
split our sample by sector and groups in order to control
for a possible industry effect. The application of our linear
model for each sector/group shows the following results:
The influence of market premium RM-RF is still the highest significant factor, and the other three factors show their
previous significance over the period. For Switzerland,
the client-facing groups show a positive significant factor toward human rights. Banks and financial firms seems
positively influenced by society and negatively by labor
changes, while the rest of the industry seems oriented
toward economic ESG news. For the UK, oil and gas shows
a highly significant positive factor for environment news,
which links to the oil split affair. For banks, there seems to
be a negative link toward society, while media has a very
positive one. The travel industry seems to have a negative
link with labor.
For the US, we find that financial services seem negatively influenced by society score changes, while oil and
gas are neutral toward such changes. Retail seems influenced negatively by economic changes and telecom by
labor changes. Technology, however, seems positively
influenced by product changes and telecom by governance
and economic.
Model 2 Analysis
The functions obtained with a non-parametric kernel
regression for each parameter over the whole Swiss sample
are as follows:
•f1 = positive linear function of RM-RF, a confirmation
or a consequence of CAPM
•f3 = positive linear function of HML with almost flat
slope
•f5 = the function of ESG Global Ethical quote score
seems to be flat until a certain amount of positive change
in score. It then becomes positive linear but with a cap,
i.e., past a certain threshold, the ESG score has little
additional influence on market performance.
ESG Impact on Market Performance of Firms: International Evidence
57
Table 7
Regression results for Model 1 - Sub-scores per year
Performance Measurement Model for January 2007 to December 2011. Stock’s return is log-return computed from the monthly market
observations of the close price from Telekurs. RF is the three -month T-bill return for UK and US, and the call money rate for Switzerland.
RM-RF is the excess return on Fama and French’s (1993) market proxy. SMB and HML are Fama and French’s factors for size and value.
MOM is the Carhart’s factor for momentum. Our ESG factors represent the variation in the news-based score on environmental, social/
societal and governance criteria. The seven “<category> Chng” (for instance ESG A_Governance.Chng) factors represent the changes
in a score computed only in one of the seven GRI dimension measured by Covalence.
StockReturn-RF ~ RM-RF + SMB + HML + MOM + ESG A_Governance Chng + ESG B_Economic.Chng + ESG C_Environment.
Chng + ESG D_Labor.Chng + ESG E_Human.Rights.Chng + ESG F_Society.Chng + ESG G_Product.Chng
CH
2007
2008
Estimate
Estimate
RM-RF
1.18 ***
0.96 ***
SMB
-0.02
-0.55 .
HML
0.24
0.03
MOM
-0.44
-0.09
ESG A_Governance.Chng
-0.001
-0.025 .
ESG B_Economic.Chng
0.009
0.047
ESG C_Environment.Chng
0.004
-0.055
ESG D_Labor.Chng
0.036
-0.005 *
ESG E_Human.Rights.Chng
-0.008
0.009
ESG F_Society.Chng
0.001
0.018 *
ESG G_Product.Chng
0.025
0.116 .
(Intercept)
-0.01 -0.02 Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1
2009
Estimate
0.98 *
-0.37
0.12
-0.40
-0.004
0.086 .
-0.068
0.006
0.003
-0.005
-0.059
0.00 2010
Estimate
1.49 ***
-0.19
0.51 .
0.12
-0.019
-0.023
-0.032
-0.012
0.001
0.012
-0.010
0.00 UK
2007
2008
Estimate
Estimate
RM-RF
1.18 ***
0.83 ***
SMB
-0.19
0.04
HML
-0.53
-0.02
MOM
-0.16
-0.19 .
ESG A_Governance.Chng
-0.015
-0.004
ESG B_Economic.Chng
-0.016 .
-0.001
ESG C_Environment.Chng
0.007
0.000
ESG D_Labor.Chng
0.002
0.001
ESG E_Human.Rights.Chng
0.002
0.006
ESG F_Society.Chng
0.020
0.000
ESG G_Product.Chng
-0.002
0.000
(Intercept)
-0.01 0.00 Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1
2009
Estimate
1.01
0.10
-0.29
-0.30
-0.014
0.000
-0.014
0.000
0.014
-0.017
-0.001
-0.01
2010
Estimate
1.14
0.05
-0.62
0.05
0.002
0.029
0.013
0.001
-0.001
0.000
0.017
0.00
US
RM-RF
SMB
HML
MOM
ESG A_Governance.Chng
ESG B_Economic.Chng
ESG C_Environment.Chng
ESG D_Labor.Chng
ESG E_Human.Rights.Chng
ESG F_Society.Chng
ESG G_Product.Chng
(Intercept)
2009
Estimate
0.92
0.10
0.10
-0.17
0.000
0.000
-0.006
0.001
0.000
0.002
-0.008
0.00
2007
Estimate
0.94
-0.08
-0.17
-0.12
0.000
0.001
0.000
0.002
-0.001
0.000
0.001
0.00
***
.
**
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1
2008
Estimate
1.10 ***
0.05
0.16 *
-0.02
-0.001
0.002
0.000
0.000
0.000
-0.002
0.000
0.00 ***
*
***
***
***
**
2011
Estimate
0.95
-0.51
0.66
-0.20
0.000
0.023
0.085
0.014
-0.031
-0.019
0.004
-0.01
***
*
*
***
.
*
2010
Estimate
1.00 ***
0.01
0.08
0.04
0.003
0.001
0.002
-0.001
0.000
-0.001
0.000
0.00 2011
Estimate
0.90 ***
0.08
0.14
-0.30
0.000
0.001
0.004
0.001
0.001
-0.001
0.001
0.00 58
Management international / International Management / Gestión Internacional, 19 (2)
Table 8
Regression results for Model 1 – Sectors and Sector’s groups
Performance Measurement Model for January 2007 to December 2011. Stock’s return is log-return computed from the monthly market
observations of the close price from Telekurs. RF is the three -month T-bill return for UK and US, and the call money rate for Switzerland.
RM-RF is the excess return on Fama and French’s (1993) market proxy. SMB and HML are Fama and French’s factors for size and value.
MOM is the Carhart’s factor for momentum. Our ESG factors represent the variation in the news-based score on environmental, social/
societal and governance criteria. The seven “<category> Chng” ( for instance ESG A_Governance.Chng) factors represent the changes
in a score computed only in one of the seven GRI dimension measured by Covalence. In this regression, UK and US firms where divided
using ICB super-sectors. In Switzerland, as the number of sample was too small, we grouped the ICB super-sectors in three custom groups
by type of activity: Consumer facing, Bank and Insurance, Industry and other.
StockReturn-RF ~ RM-RF + SMB + HML + MOM + A_Governance Chng + B_Economic.Chng + C_Environment.Chng + D_Labor.
Chng + E_Human.Rights.Chng + F_Society.Chng + G_Product.Chng
Group I
Group II
Group III
CH -by sectors
Personal & Household
Banks
Industrial Goods & Services
Goods
Insurance
Construction & Materials
Food & Beverages
Financial Services Health Care
Chemicals
Estimate
Estimate
Estimate
RM-RF
1.04 ***
1.32 ***
1.01 ***
SMB
-0.25
-0.42 .
-0.18
HML
-0.19
1.52 ***
0.01
MOM
0.24
-0.88 ***
0.13
A_Governance.Chng
-0.025
-0.017
-0.018 *
B_Economic.Chng
0.014
0.001
0.085 ***
C_Environment.Chng
-0.008
0.083
0.001
D_Labor.Chng
0.035
-0.005 **
0.005
E_Human.Rights.Chng
0.210 *
0.029
0.001
F_Society.Chng
0.003
0.018 *
-0.002
G_Product.Chng
0.038
0.002
0.002
(Intercept)
0.00 -0.01 -0.01 .
UK - by sectors
RM.RF
SMB
HML
MOM
A_Governance.Chng
B_Economic.Chng
C_Environment.Chng
D_Labor.Chng
E_Human.Rights.Chng
F_Society.Chng
G_Product.Chng
(Intercept)
RM.RF
SMB
HML
MOM
A_Governance.Chng
B_Economic.Chng
C_Environment.Chng
D_Labor.Chng
Automobiles &
Parts
Banks
Chemicals
Food & Beverages
Estimate
3.05 **
2.47 *
-2.35
-0.16
NA
-0.008
0.004
-0.126
NA
NA
NA
-0.01 Estimate
0.99 ***
0.31
0.64
-0.85
-0.004
0.050
-0.001
0.007
0.002
-0.010
0.052
0.00 Estimate
Estimate
Estimate
0.86 ***
0.66 ***
-0.19
-0.57 ***
-0.37
-0.66 *
-0.17
-0.32 **
0.017
0.002
0.055
0.061
-0.010
0.085
0.004
-0.002
-0.001
0.003
-0.013
0.081
-0.029
0.002
0.01 0.00 Estimate
Insurance
Media
Telecommunication
Utilities
Estimate
0.66 .
0.41
1.06
-0.30
1.195
-0.972
-1.159
0.017
Estimate
0.78
0.13
-0.47
0.06
-0.543
0.367
NA
-0.007
Personal &
Household Goods
***
**
***
**
Estimate
0.64 ***
0.05
-0.51 *
0.01
-0.022
0.002
-0.006
0.000
Estimate
0.70 ***
-0.10
0.69 .
0.10
-0.301
-0.162
0.460
-0.001
Health Care
Travel
& Leisure
Estimate
0.89 ***
-0.03
-0.71
-0.17
-0.005
-0.008
-0.012 .
-0.021 ***
Industrial
Goods &
Services
1.04
0.23
-0.90
0.05
0.070
0.000
-0.014
-0.012
NA
0.797
0.028
-0.01
Estimate
0.58
-0.17
-0.04
-0.03
NA
-0.068
-0.063
0.050
ESG Impact on Market Performance of Firms: International Evidence
E_Human.Rights.Chng
NA
-0.049
F_Society.Chng
0.252
0.241
G_Product.Chng
-0.218 .
-0.018
(Intercept)
0.00 0.00 Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1
Automobiles
& Parts
US - by sectors
RM-RF
SMB
HML
MOM
A_Governance.Chng
B_Economic.Chng
C_Environment.Chng
D_Labor.Chng
E_Human.Rights.Chng
F_Society.Chng
G_Product.Chng
(Intercept)
Estimate
1.42
0.76
0.22
-0.84
0.047
-0.018
-0.011
0.002
0.010
0.060
0.018
0.01
RM.RF
SMB
HML
MOM
A_Governance.Chng
B_Economic.Chng
C_Environment.Chng
D_Labor.Chng
E_Human.Rights.Chng
F_Society.Chng
G_Product.Chng
(Intercept)
Estimate
1.11
-0.34
0.68
-0.37
-0.028
0.339
-0.014
0.121
0.060
0.005
0.016
0.01
0.005
0.000
0.000
0.00 Banks
Chemicals
Estimate
1.14
-1.34
2.23
***
-0.46
0.000
0.187
0.034
0.000
0.003
-0.006
-0.005
0.01
Estimate
1.00
0.42
*
0.26
*
-0.41
-0.021
0.046
-0.001
0.003
-0.036
0.011
-0.001
0.00
***
Insurance
***
Media
***
***
***
Estimate
1.22
0.32
0.04
0.03
0.008
0.004
0.000
0.004
-0.001
0.005
0.000
0.00
59
***
***
.
*
-0.015
0.257
0.059
0.00 0.005
-0.001
0.000
0.01 Food & Beverages
Health Care
Estimate
Estimate
0.61 ***
0.88
-0.34 **
-0.45
-0.02
-0.16
***
0.05
0.04
-0.001
0.001
0.001
-0.019
0.000
-0.005
0.001
-0.001
0.000
0.001
0.001
-0.002
0.001
0.001
0.00
0.00
NA
0.266
0.054
0.00
Industrial
Goods &
Services
Estimate
***
.
***
***
.
.
1.05
-0.08
0.16
-0.13
0.001
0.001
0.001
0.001
0.000
0.000
0.002
0.00
Personal &
Household Goods
Telecommunication
Travel & Leisure
Utilities
Estimate
0.75 ***
0.01
0.29 **
-0.02
0.003
-0.003
0.001
0.000
-0.003
0.003
0.001
0.00 Estimate
1.05
-0.15
-0.62
0.07
0.013
0.051
-0.001
-0.025
-0.010
0.001
-0.013
0.00
Estimate
0.69
1.35
0.52
-0.64
0.001
-0.024
0.000
0.015
0.000
0.000
-0.004
0.01
Estimate
***
** * * * ***
***
* ***
.
.
0.73
-0.30
-0.36
0.10
0.000
0.002
0.000
-0.004
-0.003
-0.013
0.000
-0.01
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1
This could mean that ESG-related information is of
importance to investors but that investors may be unable to
distinguish between “virtuous” companies and those that
are “very virtuous.”
Non-parametric regression over the ESG score in each
category shows highly nonlinear functions for B_Economic
and C_Labor score changes, which could require further
confirmation on bigger sample or different markets.
The functions obtained with a non-parametric kernel
regression for each parameter over the whole sample are
presented in Figure 2 below. The shape of the function displayed for each ESG factors does not seems significant.
Since the non-parametric regression is sensitive to the
bandwidth, a more detailed regression could be conducted
using a non-automatic bandwidth to better tailor the variation of the data sample.
Conclusion
Our research question was how the individual company’s
market and financial performance relate to ESG criteria. We
tried to identify the influence of ESG ratings on a firm’s
market performance in Switzerland, the UK, and the US,
with two linear and nonlinear models.
In theory, a good ESG rating should signal firms with
lower residual risks and therefore increase their market
value as demand and valuation would adjust accordingly.
We tested monthly stock’s excess performance over a fiveyear period for several Swiss, US, and UK companies and
their related news-based ratings in various ESG categories.
We find a neutral or slightly negative relationship with the
overall rating for the UK but not for the US or Switzerland.
Our results regarding the sub-categories scores highlight the
fact that the link with such scores and market performance
is highly dependent on the year and sector. Those results
could be a sign that investors do not recognize ESG ratings
60
Management international / International Management / Gestión Internacional, 19 (2)
FIGURE 1
Regression results for Model 2
-0.10
-0.05
0.00
0.05
0.10
-0.10 0.00 0.10
StockReturn.RF
-0.10 0.00 0.10
StockReturn.RF
Non-linear functions for the Performance Measurement Model for January 2007 to December 2011. Model 2 is a non-linear Market
model to explain StockReturn-RF based on functions of the variables hereafter Stock’s return is log-return computed from the monthly
market observations of the close price from Telekurs. RF is the three -month T-bill return for UK and US, and the call money rate for
Switzerland. RM-RF is the excess return on Fama and French’s (1993) market proxy. SMB and HML are Fama and French’s factors
for size and value. MOM is the Carhart’s factor for momentum. Global Chng factor represent the variation in the news-based score on
environmental, social/societal and governance criteria. The graphs represent the estimated functions of the StocksReturn-RF(y-axis)
depending of the variable in the x-axis.
-0.10
0.00
-0.10 0.00 0.10
StockReturn.RF
HML
0.00
-0.10 0.00 0.10
-0.04 -0.02 0.00 0.02 0.04
-5
0.05
SMB
StockReturn.RF
-0.10 0.00 0.10
StockReturn.RF
RM.RF
-0.15
-0.10
-0.05
0.00
0.05
MOM
5
10
Global.chng
variation as a flag of a lower/higher residual risk, except
for periods where the market is sensitive to specific conditions. Only under those conditions would the prices adjust
to the better/worst perception of the risk of the firm, which
could be an interesting topic to expand in the field of behavioral finance. We also consistently measured the impact of
ESG news-based ratings on the stock’s market return to
be smaller than the Fama-French and momentum factors.
Our results should, however, be considered with care as
our sample only consists of hundreds of firms and as such,
should be extended to a larger number of firms and a longer
observation period in order to confirm the link with theory.
The kernel regression for Switzerland displays a nonlinear
relation for news-based ratings toward the market over the
whole period, which could be taken into account and may
lead to further studies using a non-linear relationship.
To conclude, the studies following the stakeholder theory (Freeman, 1984) postulates that there are some benefits
for firms to improve their ESG ratings as this could increase
their performance. However, we show that this link has yet
to be fully understood and recognized by the market, as it
will not sanction an overall monthly increase or decrease of
ESG ratings, except during specific, contextual periods. It
is an interesting result for a firm’s management, who might
want to show their good deeds in periods when this factor is
under exposure -for instance, when there is a discussion on
a new regulation that the firm is already compliant with. It
is also interesting for public policy makers and regulators to
know that the market does not clearly sanction negative or
positive ESG efforts yet and that firms or investors, despite
being favorably minded toward sustainability, might need
further incentives from them.
ESG Impact on Market Performance of Firms: International Evidence
61
FIGURE 2
Regression results for Model 2 – Sub-scores
0.00
0.10
StockReturn.RF
-0.05
0.00
0.05
10
-2
-1
-20
1
2
3
4
10
-40
-20
0
10
E_Human.Rights.Chng
0.15
StockReturn.RF
10
-0.10
0
F_Society.Chng
0
0
B_Economic.Chng
D_Labor.Chng
0.15
StockReturn.RF
-0.10
-10
6
StockReturn.RF
-40
C_Environment.Chng
-20
4
0.15
StockReturn.RF
5
-0.10
0.00
2
A_Governance.Chng
0.15
StockReturn.RF
-0.10
-0.10
0
0.15
StockReturn.RF
-2
-0.10
0.00
MOM
0.15
-0.05
-0.10
-0.10
-0.04 -0.02 0.00 0.02 0.04
HML
0.15
StockReturn.RF
-0.15
0.05
SMB
-0.10
StockReturn.RF
-0.10 0.00 0.15
RM.RF
-0.10 0.00 0.15
StockReturn.RF
-0.10
-0.10 0.00 0.15
StockReturn.RF
-0.10 0.00 0.15
Non-linear functions for the Performance Measurement Model for January 2007 to December 2011. Model 2 is a non-linear Market
model to explain StockReturn-RF based on functions of the variables hereafter Stock’s return is log-return computed from the monthly
market observations of the close price from Telekurs. RF is the three -month T-bill return for UK and US, and the call money rate for
Switzerland. RM-RF is the excess return on Fama and French’s (1993) market proxy. SMB and HML are Fama and French’s factors
for size and value. MOM is the Carhart’s factor for momentum. Global Chng factor represent the variation in the news-based score on
environmental, social/societal and governance criteria. The graphs represent the estimated functions of the StocksReturn-RF(y-axis)
depending of the variable in the x-axis.
-8
-6
-4
-2
0
2
G_Product.Chng
This study also contributes to the literature evaluating
the relationship between financial and CSR performances,
and the non-parametric response of performance to ESG
criteria may open a new way of research to better understand the complexity of this relationship (Orlitzky; 2013).
In addition, it would be interesting to further study the link
between an ESG news-based rating and market performance with regard to a larger sample and other countries,
as well as study the link between those returns and financial performance using accounting models over the same
period.
Bibliographie
Albuquerque, R., Durnev, A., Koskinen, Y., 2012. Corporate
Social Responsibility and Asset Pricing in Industry
Equilibrium. Working paper.
Angel, J.J., Rivoli, P., 1997. Does ethical investing impose a cost
upon the firm? The Journal of Investing 6 (4), 57-61.
Balakrishnan, R. Sprinkle, G.B., Williamson M.G., 2011.
Contracting Benefits of Corporate Giving. An Experimental
Investigation. The Accounting Review 86 (6), 1887–1907.
Banz, R.W., 1981. The relationship between return and market value
of common stocks. Journal of Financial Economics 9, 3–18.
62
Management international / International Management / Gestión Internacional, 19 (2)
Battacharya, C.B., Sen, S., Korschun, D., 2008, Using Corporate
Social Responsibility to Win the War for Talents. MIT Sloan
Management Review 49 (2), 37–44.
Jensen, M.C., 1986. Agency Costs of Free Cash Flow, Corporate
Finance, and Takeovers.American Economic Review 76, 323.
Capelle-Blancard, G. and Monjon S., 2011. The performance of
socially responsible funds does the screening process matter?
CEPII, Working paper n°2011-12.
K ang, K.H., Lee, S., Huh, C., 2010. Impacts of positive and negative corporate social responsibility activities on company performance in the hospitality industry. International Journal of
Hospitality Management 29, 72–82.
Carhart, M.M., 1997. On Persistence in Mutual Fund
Performance. The Journal of Finance 52(1), 57–82.
Kurtz, L., 2005. Answers to Four Questions. Journal of Investing
14(3), 125–139.
Edmans, A., 2011. Does the stock market fully value intangibles?
Employee satisfaction and equity prices. Journal of Financial
Economics 101 (3), 621–640.
Mackey, A., Mackey, T.B., Barney, J.B., 2007. Corporate social
responsibility and firm performance: investor preferences and
corporate strategies. Academy of Management Review 32 (3),
817–835.
Eccles, R.G., Ioannou, I., Serafeim, G., 2012. The Impact of a
Corporate Culture of Sustainability on Corporate Behavior
and Performance. National Bureau of Economic Research
Working Paper Series n°17950.
Economist, 2008. Just good business. Special report on CSR.
January, 19th.
El Ghoul, S., Guedhami, O., Kwok, C.C.Y., Mishra, D.R., 2011.
Does corporate social responsibility affect the cost of capital?
Journal of Banking & Finance 35(9), 2388–2406.
Fama, Eugene F., K enneth R. French, 1993. Common risk factors in the returns on bonds and stocks. Journal of Financial
Economics 33 (1), 3–53.
Freeman, R.E., 1984. Strategic management: A stakeholder
approach. Cambridge University Press, Cambridge.
Margolis, J., Elfenbein, H., Walsh, J., 2007. Does it pay to be
good ? A meta-analysis and redirection of research on the
relationship between corporate social and financial performance, Working Paper Harvard University.
McWilliams, A., Siegel, D., 1997. Event studies in management research: theoretical and empirical issues. Academy of
Management Journal 40 (3), 626–657.
Myers, S.C., 1977. Determinants of Corporate borrowing. Journal
of Financial Economics 5, 147–175.
Ohlson, J.A., 1995. Earnings, Book Values, and Dividends in
Equity Valuation. Contemporary Accounting Research 11 (2),
661–687.
Friedman, M., 1970. The Social Responsibility of Business is to
Increase its Profits. New York Times Magazine.
Orlitzky, M., Schmidt, F.L., Rynes, S.L., 2003.Corporate Social
and Financial Performance: A Meta-Analysis. Organization
Studies, 24 (3), 403–441.
Galema, R. Plantinga, A. and B.Scholtens. 2008. The stocks
at stake: Return and risk in socially responsible investment.
Journal of Banking & Finance 32 (12), 2646–2654.
Orlitzky, M., 2013. Corporate Social Responsibility, Noise,
and Stock Market Volatility. Academy of Management
Perspectives 27 (3), 238–254.
Garriga, E., Melé, D., 2004. Corporate Social Responsibility
Theories: Mapping the Territory. Journal of Business Ethics
53, 51–71.
Porter, M.E., 1991. America’s green strategy. Scientific
American, 264 (4), 168.
Gompers, P., Ishii, J., Metrick, A., 2003. Corporate Governance
and Equity Prices. Quarterly Journal of Economics 118 (1),
107-156.
Graff Zivin, J., Small, A., 2005. A Modigliani-Miller Theory
of Altruistic Corporate Social Responsibility. B.E. Journal of
Economic Analysis & Policy: Topics in Economic Analysis &
Policy 5, 1–21.
Gramlich, D. Finster, N., 2013; Corporate Sustainability and
Risk. Journal of Business Economics 83, 631–664.
Gregory Birth, Laura Illia, Francesco Lurati, Alessandra
Zamparini, 2008. Communicating CSR: practices among
Switzerland’s top 300 companies. Corporate Communications:
An International Journal 13, 182–196.
Hong, H., K acperczyk, M., 2009. The price of sin: The effects
of social norms on markets. Journal of Financial Economics
93, 15–36.
Horváthová, E., 2010. Does environmental performance
affect financial performance? A meta-analysis. Ecological
Economics 70, 52–59.
Ionnou, I., Serafeim, G., 2011. The Consequences of Mandatory
Corporate Sustainability Reporting. Working Paper, Harvard
Business School.
Jensen, M.C., Meckling, W.H., 1976. Theory of the firm:
Managerial behavior, agency costs and ownership structure.
Journal of Financial Economics 3, 305–360.
Porter, M.E., K ramer, M.R., 2011. The big idea: created shared
value. Harvard Business Review, January-February, 4-17.
R enneboog, L., Horst, T., Zhang, C., 2008. Socially responsible
investments: Institutional aspects, performance, and investor
behavior. Journal of Banking & Finance, 32 (9), 1723–1742.
R evelli, C., Viviani, J, 2013. Performance financière de
l’investissement socialement responsable (ISR) : une métaanalyse. Finance Contrôle Stratégie [Online], 15-4 |2013,
Online since 18 March 2013. URL : http://fcs.revues.org/1222.
Russo, M.V., Fouts, P.A., 1997. A resource-based perspective
on corporate environmental performance and profitability.
Academy of Management Journal 40 (3), 534–559.
Sandberg, J., C., Juravle, T.D. Hedesström, and I. Hamilton,
2009. The heterogeneity of socially responsible investment.
Journal of Business Ethics, 87, 519-533.
Sharfman, M.P., Fernando, C.S., 2008. Environmental risk
management and the cost of capital. Strategic Management
Journal 29 (6), 569–592.
Statman, M., Glushkov, D., 2009. The Wages of Social
Responsibility. Financial Analysts Journal 65 (4), 33–46.
Teoh, S.H., Welch, I., Wazzan, C.P., 1999. The Effect of Socially
Activist Investment Policies on the Financial Markets: Evidence
from the South African Boycott. Journal of Business 72, 35–89.
Waddock, S.A., Graves, S.B., 1997. Finding the link between stakeholder relations and quality. Journal of Investing 6 (4), 20-24.
ESG Impact on Market Performance of Firms: International Evidence
63
Appendix A
Fig. A1 - Socially Responsible Investment – Acknowledged Strategies
Type
Strategies
Definition
Negative
Screening
Exclusion
Exclusion of certain sectors such as weapons etc.
Norm-based screening
Exclusion based on compliance with international standard and norms
Positive
screening
ESG Integration
Integration of ESG criteria to classic Financial analysis
Best-in-Class
Selection or Weighting of stocks according to ESG criteria
Themed Funds
Funds with a theme focused on Sustainability , e.g. Green energy, Health, etc.
Engagement Voting
Active Ownership through share voting on ESG topics
Impact Investment
Investing for a clear ESG impact e.g. Microfinance, local business funds, etc...
Active
Investment
Fig. A2 - US and European SRI Growth –US SIF 2012 Executive Summary report, EURO SIF 2012 report
SRI in the US IN $bn
SRI in Europe in €bn
1997
1997
1999
2003
2005
2007
2010
2012
639
1’185
2’159
2’323
2’290
2’711
2’069
3’744
2005
2007
2009
2011
1’768
4’066
*7’375
*11’661
* includes norm-based screening since 2009 - 2009 988bn-2011 2’346bn
Документ
Категория
Без категории
Просмотров
0
Размер файла
424 Кб
Теги
1030386ar
1/--страниц
Пожаловаться на содержимое документа